Now Accepting Apple Pay

Apple Pay is the easiest and most secure way to pay on StudyMoose in Safari.

Biometrics Identification

Abstract

The face recognition has become acommonmeans of identity authentication because of the advantagesof uniqueness, non-invasive and not easy to be stolen. Inmodern times, face recognition has become one of the keyaspects of computer vision. The outsourcing of facerecognition to the service provider is a typical mannernowadays. There are many biometric processes, in that facerecognition is the best method. In this paper we are using.

Convolutional Neural Network (CNN) for face featureextraction and for the authentication purpose we make useof secure nearest neighbor algorithm[5].

Key Words :Face Recognition, Convolutional NeuralNetwork (CNN), Face Verification, Secure NearestNeighbor.

INTRODUCTION

In recent years, biometrics identification such as the face, iris, fingerprints and DNA receives significant attention, especially in the field of human identi fiction and authentication. Compared with traditional password-based authentication, biometrics has the advantages of uniqueness, distinctive, mobility, user-friendliness and not being transferable. Face verification, as the most popular technology in biometrics, is widely used in authentication systems because it’s non-invasive and not needing the cooperation of users when scanning face images compared with the identification technology of fingerprint and iris.

Get quality help now
Bella Hamilton
Verified writer

Proficient in: Cloud Computing

5 (234)

“ Very organized ,I enjoyed and Loved every bit of our professional interaction ”

+84 relevant experts are online
Hire writer

However, in addition to the advantages mentioned above, there are also many challenges to the technology of biometric identification such as privacy and security. It is a very typical choice to perform face verification by the cloud service providers such as face++, but in this situation, one must upload the face image to the cloud server of service providers. Face image is extremely sensitive information, for itcontains much privacy. In this paper, we propose a frame of identity authentication based on the technology of face recognition in which a secure nearest neighbor scheme is used for face recognition and convolutional neural network (CNN) isused for face feature extraction.

Get to Know The Price Estimate For Your Paper
Topic
Number of pages
Email Invalid email

By clicking “Check Writers’ Offers”, you agree to our terms of service and privacy policy. We’ll occasionally send you promo and account related email

"You must agree to out terms of services and privacy policy"
Check writers' offers

You won’t be charged yet!

RELATED WORK

A Secure Protocol for Biometric Identification

In recent years, many privacy-preserving methods for biometric data recognition were proposed. The combination of cryptographic primitives such as homomorphic encryption and garbled circuits are mostly used to protect the biometric data in these methods.A secure protocol for privacy preserving biometric identification was proposed in [1] that can achieve security against semi-honest adversaries. The security model is designed by the scheme of homomorphic encryption, oblivious transfer and garbled circuit. Hamming distance and Euclidean distance are used for measuring the similarity of biometric information.

A Protocol for Outsourcing

Identification of encrypted Biometric DataIn [2], a protocol for outsourcing identification of encrypted biometric data to the untrusted server was proposed, anew method of oblivious RAM was adopted for iris recognition and proved to be effective. The proposal only rely on standard symmetric encryption technology. The result shows that itis the only one that can deal with large databases and utilize all the opportunities of cloud computing

Finger and Iris-Recognition Based User Authentication

A privacy-preserving fingerprint recognition scheme was given in [3] that improved the performance of computation and communication efficiency. In [4], the privacy-preserving computation method was further improved that can be used for computing Euclidean distance, Hamming distance, Mahalanobis distance and scalar product of different biometric traits such as iris, face, and fingerprint.

System design is the process of defining the architecture, components, modules, interfaces, and data for asystem to satisfy specified requirements. System design could be seen as the application of systems theory to product development. The proposed user authentication systemcan be divided into four main modules. The modules and their functions are defined in this section. The four modules into which the proposed system isdivided are:

Image Capture

We make use of WEB camera to capture the frontal images of the user and the further process goes for face detection.

 Face Identification

Face verification isthe task of determine whether two face images belong to the same person. Face feature extraction is the first step of face verification. In early researches, many classical methods appeared, such as SIFTs, LBPs, and Gabor features. In recent years, CNNs-based methods have proven to be more efficient for face feature extraction. The structure of network in DeepID is shown in Figure 1. The network contains four convolutional layers and four max-pooling layers. Through this network, a160-dimensional feature vectors can be extracted from aface image. In this paper, method of DeepID is used to extract the facial feature vector.

Figure 1: Face Feature Extraction

Storing the Extracted Image Data in Database

As we chose biometric based system, enrolment of every individual is required. This database development phase consists of image capture of every individual and extracting the biometric feature, in our case itisface, and later it is enhanced using DeepID techniques and stored inthe database.

User Authentication

Figure 2: Flow Chart of Proposed SystemFace feature vector can reveal some information of face image content. Therefore, the plain text of the face feature vector can not be stored without protecting, an effective encryption method must be employed to protect the plaintext of the face feature vector from being stolen adversary. In this paper, the secure nearest neighbor algorithm is used as our encryption method that can not only protect the privacy of face feature vector but also guarantee the ciphertext of the feature vectors can be used for face recognition. The most obvious advantage of the secure nearest neighbor algorithm is that it doesn’t need high computation complexities or communication burden compared with the homomorphic encryption algorithm. On getting the face image, the CNN will firstly extract a feature vector from it, then perform the secure nearest neighbor algorithm to encrypt the feature and store the ciphertext in its database-1. The database-1 of each computing server is used for storing the face information of the users who registered on it. For the identity authentication, the system searches in its database-1 using the method of secure nearest neighbor to find whether the user is registered on it, ifso, we can get his permission immediately. Identity authentication can be completed if the user’s feature vector is found in the database-1 of the system. If the feature vector does not match then the user authentication fails.

EFFICIENCY EVALUATION

We compare the time consumption of face recognition achieved by the plaintext of face data with the ciphertext encrypted by the privacy-preserving method we proposed above. The time consumption on the method of secure nearest neighbor contains the time of feature vectors extraction, encryption and recognition. The results about time consumption of these two situations are recorded according to considerable experiments. We can find that the time consumption of these two methods is almost linear to the number of face images in the database from our experimental results. Notice that, the time consumption of face recognition with the face feature vectors protected by the secure nearest neighbor is almost equal to that with the plain text of face feature vectors. Thus, performing a secure nearest neighbor scheme on the system can protect the face data of the users registered unit from being stolen by the adversary and won’t waste too much time at the same time.

CONCLUSIONS

A convolutional neural network-based face verification system for user authentication is proposed in this paper. Face feature vectors are extracted by the method of convolutional neural networks. Secure Nearest NeighborMethod is introduced in our system to perform face verification. All operations in the system is preformed on the encrypted feature vectors to prevent privacy leaks. How to ameliorate the encryption algorithm to further reduce the time consumption of authentication is still an open problem and will continue to be studied in our future work.

ACKNOWLEDGEMENT

With all respect and gratitude, we would like to thank all the people who have helped us directly or indirectly for the completion of the project “A Convolutional NeuralNetwork-Based Face Verification System for UserAuthentication “. We express our hearty gratitude towards Prof. Pallavi Gowdoor for guiding us to understand the work conceptually and also for her constant encouragement to complete the paper. Our association with her as a student has been extremely inspiring. We would like to give our sincere thanks to Dr. Hemalatha K.L. Head of the Department of information science and Engineering for her technical support and constant encouragement. We would also like to extend our sincere thanks to our Principal Dr. Manjunatha A. for his help and support in all respects. We would also like to thank all our staff members and collogues who helped us directly or indirectly throughout our dissertation work.

REFERENCES

  1. M. Blanton and P. Gasti, ”Secure and ef?cient protocols for iris and fingerprint identification, presented at the Eur. Symp. Res. Comput. Secur., Leuven, Belgium, Sep. 2011.
  2. J. Bringer, H. Chabanne, and A. Patey, ”Practical identification with encrypted biometric data using oblivious ram, ”presented at the Int. Conf. Biometrics (ICB), Madrid, Spain, Jun. 2013.
  3. A.-R. Sadeghi, T. Schneider, and I. Wehrenberg, ”Ef?cient privacy-preserving face recognition, ” presented at the Int. Conf. Inf. Secur. Cryptol., Seoul, South Korea, Dec. 2009.
  4. J.Bringer,H. Chabanne, M.Favre, A.Patey, T.Schneider, and M.Zohner, ”GSHADE: Faster privacy-preserving distance computation and biometric identi ?cation, ” presented at the 2nd ACM WorkshopInf. Hiding Multimedia Secur., Salzburg, Austria, Jun. 2014.
  5. W. K. Wong, D. W.-L. Cheung, B. Kao, and N. Mamoulis, ”Secure knn computation on encrypted databases, ” presented at the ACM SIGMOD Int. Conf. Manage. Data, New York, NY, USA, Jun./Jul. 2009.

Cite this page

Biometrics Identification. (2019, Dec 14). Retrieved from http://studymoose.com/biometrics-identification-essay

👋 Hi! I’m your smart assistant Amy!

Don’t know where to start? Type your requirements and I’ll connect you to an academic expert within 3 minutes.

get help with your assignment